Austria asks the EU to consider giving Anthropic a home in Europe
June 28, 2026
A European government has, in writing, asked the European Union to find a way to host an American artificial-intelligence company. The request is unusual enough that the official making it conceded, in the same breath, that people would doubt it could be done.
The official is Alexander Pröll, Austria’s state secretary for digitalisation, and the company is Anthropic. In a letter to Henna Virkkunen, the European Commission’s executive vice-president for tech sovereignty, Pröll urged member states to explore “the strategic establishment and participation of Anthropic within the European Union,” according to Bloomberg, which first reported the letter.
Pröll’s pitch is that Europe should not simply absorb US export controls on frontier AI models. “Let us jointly explore the strategic establishment and participation of Anthropic within the European Union,” he wrote, “with legal certainty, market access, capital and a set of values that suits this company.”
The move follows a June 2026 US Commerce Department directive that forced Anthropic to suspend foreign access to its advanced Mythos 5 and Fable 5 models due to national security concerns. The directive blocked access by any foreign national, whether inside or outside the United States.
Pröll argued that Europe must secure direct access to frontier AI to avoid dependence on US policy decisions. He did not specify whether Austria envisages a European subsidiary, a data-residency arrangement, an equity stake, or something looser, and acknowledged there would be scepticism about whether the idea was feasible at all.
What he insisted on was the stakes: that Europe must not be cut off from major innovations, and that being a customer is not the same as having a guarantee.
Austria’s letter is the same anxiety, pushed one step further: rather than negotiate access, secure presence.
Whether the Commission entertains the proposal is an open question. Virkkunen’s office had not publicly responded to the letter, and any move of this kind would involve legal, competition, and security questions that no single member state can settle.
Context: On June 27, 2026, Anthropic received partial US government approval to redeploy Mythos 5 to a limited group of critical infrastructure operators. OpenAI simultaneously limited GPT-5.6 Sol to government-approved partners. The parallel restrictions have intensified European debate over AI sovereignty and dependence on US frontier model providers.